Reducing computing times in optimizing energy system models – challenges and opportunities

Energy systems based on renewable resources are characterized by a high number of decentralized components as well as by an increasing importance of storage, grid, and demand response. State-of-the-art energy system models reflect this growing complexity by combining a comprehensive representation of energy sectors and technologies with a high spatial and temporal resolution. However, such high level of detail is typically associated with long solving times of the models. The reduction of the latter is urgently required and the need for guidelines for a reasonable balance between degree of detail and solution time is pressing. The BEAM-ME project addresses these questions by improved computing power and efficiency in energy systems modelling. The German Aerospace Center is the principal investigator on the modelling side, while the project gathers various partners with complementary expertise in the fields of optimization algorithms, high performance computing, and application development. This talk provides an overview and evaluation of conceptual and technical strategies identified so far to reduce the model solution time of the REMix energy system model. Moreover, an insight in the implementation and assessment of selected speed-up strategies is provided. Finally, conclusions from the first project results are drawn and an outlook on the subsequent works is given.